Nowadays, natural disasters occur frequently in the world, causing a large number of casualties and economic losses. As a typical natural disaster caused by non-human activities, debris flow can not be fully prevented in advance. Therefore, timely emergency rescue and relief work after the disaster is particularly important. This paper takes Yunnan Province, a province with severe debris flow disaster, as an example to discuss the formulation of a reasonable and effective debris flow rescue and relief program, so as to ensure that the rescue and relief work can be carried out in an orderly manner and avoid greater losses caused by disasters. This paper uses a combination of qualitative analysis and quantitative analysis to study the deployment of personnel and material resources in the debris flow rescue and relief plan. First, we select the five influencing factors of the 24 hours precipitation, the number of rainstorm days in the month, the precipitation and the altitude of the month, and the terrain slope when the debris flow disaster occurs. We obtain the weight of each influencing factor through the Analytic Hierarchy Process (AHP), and use Fuzzy Evaluation Method to evaluate the risk of debris flow disaster index. Then through qualitative analysis, we concluded that debris flow rescue and disaster relief requires 3 types of rescue team personnel and 6 types of emergency supplies. According to the three indicators of the debris flow risk evaluation index, the affected population, and the direct economic loss of the disaster, the rescue personnel are dealt with through the Radial Basis Function neural network (RBFNN), so as to construct a debris flow rescue and disaster relief plan. Finally, taking the mud-rock flow event in Gongshan County, Nujiang Prefecture, Yunnan Province on May 31, 2018, as an example, a relevant plan for the event was formulated according to the rescue and relief plan constructed in this paper, and the feasibility of the plan was evaluated by comparing with the actual disaster relief data of the event.